PROJECT SUMMARY Therapy-related myelodysplasia/acute myeloid leukemia (t-MDS/AML) is a lethal complication of autologous hematopoietic cell transplantation (aHCT) for Hodgkin lymphoma (HL) and non-Hodgkin lymphoma NHL). The 10-year cumulative incidence of t-MDS/AML is ~6% to 8%. Post-aHCT t-MDS/AML is associated with older age at aHCT, pre-aHCT exposure to alkylators, etoposide and radiation, peripheral blood stem cell (PBSC) mobilization with etoposide, and conditioning with total body irradiation (TBI). t-MDS/AML is characterized by poor response to conventional chemotherapy, and median survival of <10 months. t-MDS/AML is the leading cause of non-relapse mortality among aHCT recipients for HL/NHL. It is generally believed that hematopoietic stem cells (HSCs) exposed to cytotoxic therapy suffer genomic damage leading to malignant transformation. However, high inter-individual variation in t-MDS/AML risk suggests a potential role for genetic susceptibility. Previous reports (using a candidate gene approach) suggest an association between germline single nucleotide polymorphisms (SNPs) and t-MDS/AML risk. We are conducting a genome-wide association study (GWAS: Illumina HumanOmni5-Quad BeadChip platform; 303 cases; 606 controls) to identify germline variants associated with t-MDS/AML; top SNPs will contribute to creation of genetic profile. Somatic mutations in leukemia-associated genes DNMT3A, ASXL1, and TET2 seen in peripheral blood in ~10% of older healthy population, are associated with >10-fold increase in risk for subsequent leukemia; targeted next-generation sequencing will be used to identify driver mutations. We observed altered gene expression in PBSC samples from patients who subsequently developed t-MDS/AML when compared with patients who did not. This information was used to develop a 38-gene PBSC classifier in an independent test set; this classifier will also contribute to the genomic profile. The elevated risk of t-MDS/AML after aHCT, coupled with the poor prognosis, present an unmet need for pre-aHCT identification of patients at increased risk for post-aHCT t-MDS/AML to guide use of alternative therapeutic options for HL/NHL management. We hypothesize that a combined clinical and genetic risk prediction model applied prior to aHCT will allow identification of HL/NHL patients at increased risk for post-aHCT t-MDS/AML. The City of Hope cohort with the available PBSC products will serve as the Discovery cohort (n=1,915). aHCT recipients for HL/NHL with PBSC product at the University of Nebraska or University of Minnesota will be utilized as an independent Validation cohort (n=2,036). The two cohorts will be used to develop a prediction model that includes clinical predictors (therapeutic exposures, conditioning, stem cell mobilization and CD34+ cell dose) and genetic factors (SNPs, gene expression profile and clonal somatic mutations) to optimize pre-aHCT identification of HL/NHL patients at highest risk for t-MDS/AML after aHCT. This proposal is innovative in its use of a comprehensive, evidence-based approach to develop a risk prediction model for t-MDS/AML that can be adopted prior to aHCT, allowing for risk-informed personalized treatment.